The short answer is yes, and no. The long answer is, well, read on to find out.
Several have been asking the question, including myself, wondering whether Artificial Intelligence is just another passing fad like maybe the Google Glass or nano technology.
The hype for AI began over the past few years, although if you actually look back at the 60’s it seems to have started way back then. In the early 90s and all the way down to the early 2000’s, a lot of media and television shows were talking about AI quite a bit. Going 25 centuries even further back, Aristotle speaks of not just thinking machines but goes on to talk of autonomous ones in his book, Politics:
for if every instrument, at command, or from a preconception of its master's will, could accomplish its work (as the story goes of the statues of Daedalus; or what the poet tells us of the tripods of Vulcan, "that they moved of their own accord into the assembly of the gods "), the shuttle would then weave, and the lyre play of itself; nor would the architect want servants, or the [1254a] master slaves.
Aristotle, Politics: A treatise on Government, Book 1, Chapter 4
This imagery of AI has managed to sink into our subconscious minds over the centuries propelling creative work, academic research and industrial revolutions toward that goal. The thought of giving machines a mind of their own, existed quite long ago, but recent advancements in technology have made it much clearer and realistic.
The year is 2018. The 4th Industrial Revolution is happening and intelligent automation has taken over. This is the point where I say no, AI is not overhyped.
General Electric, for example, is a billion dollar manufacturing company that has already invested in AI. GE Digital has AI systems running through several automated systems. They even have their own IIoT platform called Predix.
Similarly, in the field of healthcare, the implementation of AI is growing in leaps and bounds. The Google Deepmind project is able to process millions of medical records within minutes. Although this kind of research is in its early phase, Google is working closely with the Moorfields Eye Hospital NHS Foundation Trust to implement AI and improve eye treatment. AI startups focused on healthcare and other allied areas such as genetic engineering are some of the highly invested and venture capital supported ones in recent times.
Computer Vision or image recognition is one field where AI has really proven its power. Analysing datasets like iris has never been easier, paving way for more advanced use cases like automated quality checks in manufacturing units.
Another interesting field is Healthcare, where AI has helped sift through tonnes of data, helping doctors diagnose illnesses quicker, manufacture more effective and responsive drugs, and in patient monitoring.
The list is endless, clearly showing that AI has made its mark in several industries.
Now, if you talk about the commercial implementations of AI, they’re still quite far fetched at the moment. Take the same Computer Vision application for example. Its implementation will be a huge breakthrough in autonomous vehicles. But if researchers have managed to obtain an 80% accuracy for object recognition on roads, the battle is not close to being won! Even if they do improve, do you think driverless vehicles are ready to drive in the snow, through the rain or even storms?
I remember a few years ago, Business Process Outsourcing was one industry, at least in India, that was quite fearful of the entry of AI and autonomous systems that might take over their jobs. Machines are only capable of performing 60-70% of the BPO processes in Insurance, and with changing customer requirements and simultaneously falling patience levels, these numbers are terrible!
It looks like the end of Moore’s law is here, for AI I mean. Well, you can’t really expect AI to have the same exponential growth that computers did, decades ago. There are a lot of unmet expectations in several fields, which has a considerable number of people thinking that AI isn’t going to solve their problems now, and they’re right. It is probably going to take a few more years to mature, making it a thing of the future, not of the present. Is AI overhyped now? Yeah, maybe?
Someone once said, hype is a double-edged sword. If it’s not enough, innovation may become obscure and if it’s too much, expectations will become unreasonable. It’s true that AI has several beneficial use cases, but what about fairness of such systems? Will machines continue to think the way they’re supposed to or will they start finding their own missions that don’t involve benefits to the human race? At the same time, there’s also a question of security and data privacy. GDPR will come into effect in a few days, but what about the prevailing issues of internet security?
I had an interesting discussion with a colleague yesterday. We were talking about what the impact of AI could be for us as end-customers, in a developing and young country like India. Do we really need to fear losing our jobs, will we be able to reap the benefits of AI directly or would it be an indirect impact? The answer is, probably yes, but not so soon.
If we drew up the hierarchy of needs pyramid for AI, it would look something like the above. For each field to fully leverage AI, it’s going to involve several stages like collecting data, storing it effectively, exploring it, then aggregating it, optimising it with the help of algorithms and then finally achieving AI. That’s bound to take a LOT of time!
Honestly speaking, a country like India lacks as much implementation of AI in several fields. The major customers of AI, apart from some industrial giants, will obviously be the government. Although, that is sure to take at least a decade or so, keeping in mind the several aspects to be accomplished first. In the meantime, buddying AI developers and engineers are scurrying to skill themselves up in the race to be in the cream of the crowd!
Similarly, what about the rest of the world? Well, I can’t speak for everyone, but if you ask me, AI is a really promising technology and I think we need to give it some time; allow the industries and organisations investing in it to take enough time to let it evolve and ultimately benefit us customers, one way or another.
You can now make music with AI thanks to Magenta.js
Splunk leverages AI in its monitoring tools